Skip to content
aiqbench

Qwen · model passport

Qwen: Qwen3 32B

qwen3-32b
  • Open weights
  • Apache 2.0 licence
  • 32.8B params
  • 131,072 context
  • Released 2025-04-29(curated)
11/27axes covered
Overall · LMArena
1,347#205 of 396
Best percentile
78.1HHEM Hallucination (Vectara)
Cohort standing
53.6median 48.6

What the data says

Benchmarked on 11 of 27 tracked axes. Every figure below is the model's own published result on that board.

Uneven: 46.6 percentile points between its best and worst axis — 78.1 pct on HHEM Hallucination (Vectara) versus 31.5 pct on IFBench (Artificial Analysis).

  • HHEM Hallucination (Vectara)78.1 pct94.1%Verify →
  • BFCL v4 (function calling)Dormant board72 pct48.7%Verify →
  • MMLU-Pro (Artificial Analysis)65.9 pct79.8%Verify →
  • LiveCodeBench (Artificial Analysis)60.3 pct54.6%Verify →
  • Coding51.7 pct1,406Verify →
  • Hard prompts50 pct1,367Verify →
  • Overall48.5 pct1,347Verify →
  • AIME (math)48 pct66.9%Verify →
  • Instruction following47.7 pct1,331Verify →
  • GPQA Diamond47.1 pct65.7%Verify →
  • IFBench (Artificial Analysis)31.5 pct36.3%Verify →
  • WebDev ArenaNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • LiveBenchNot benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

RepositoryQwen/Qwen3-32B
Parameters32.8B
ArchitectureQwen3ForCausalLM
LicenceApache 2.0

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Qwen/Qwen3-32B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Use it hosted

API model IDqwen/qwen3-32b
Context length131,072 tokens
Max output16,384 tokens
Input price$0.08 per 1M tokens
Output price$0.28 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogit biasLogprobsMax tokensMin-pPresence penaltyReasoningRepetition penaltyResponse formatSeedStop sequencesStructured outputsTemperatureTool choiceTool callingTop-aTop-kTop logprobsTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen/qwen3-32b","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • qwen3-32b stands at 53.6 (raw mean 54.6) — measured on 11 of 27 axes.

    standingtemplateVerify →

  • qwen3-32b stands at 53.6 versus a cohort median of 48.6 across 150 models released within 180 days — above the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →